High dimensional single-cell analysis reveals iNKT cell developmental trajectories and effector fate decision

In [1]:
import numpy as np
import pandas as pd
import scanpy as sc
import scrublet as scr
import doubletdetection
import warnings
import matplotlib.pyplot as plt
import seaborn as sns
warnings.filterwarnings('ignore')

import logging
mpl_logger = logging.getLogger('matplotlib')
mpl_logger.setLevel(logging.WARNING) 

sc.settings.verbosity = 3             # verbosity: errors (0), warnings (1), info (2), hints (3)
sc.logging.print_versions()

scanpy==1.4.4.post1 anndata==0.6.22.post1 umap==0.3.10 numpy==1.17.3 scipy==1.4.1 pandas==0.25.3 scikit-learn==0.21.3 statsmodels==0.10.2 python-igraph==0.7.1 louvain==0.6.1
In [2]:
sc.settings.set_figure_params(dpi=130)

WT

Back to top

In [3]:
adata_ann = sc.read_h5ad('./output/wt.ann.h5ad')
adata_raw = sc.read_h5ad('./output/wt.preprocessing.h5ad')
adata_raw.shape
Out[3]:
(3288, 12958)

PAGA all clusters

In [4]:
sc.tl.paga(adata_ann, groups='cell_type')

running PAGA
    finished: added
    'paga/connectivities', connectivities adjacency (adata.uns)
    'paga/connectivities_tree', connectivities subtree (adata.uns) (0:00:00)
In [5]:
sc.pl.paga(adata_ann, color=['cell_type'], threshold=0.3)

--> added 'pos', the PAGA positions (adata.uns['paga'])
In [6]:
sc.tl.draw_graph(adata_ann, init_pos='paga')
sc.pl.draw_graph(adata_ann, color=['cell_type'])

drawing single-cell graph using layout 'fa'
    finished: added
    'X_draw_graph_fa', graph_drawing coordinates (adata.obsm) (0:00:20)
In [7]:
sc.pl.paga_compare(
    adata_ann, threshold=0.3, title='', right_margin=0.2, size=10, edge_width_scale=0.5,
    legend_fontsize=0, fontsize=0, frameon=False, edges=True)

--> added 'pos', the PAGA positions (adata.uns['paga'])
In [8]:
sc.pl.paga_compare(
    adata_ann, threshold=0.35, title='', right_margin=0.2, size=10, edge_width_scale=0.4,
    legend_fontsize=0, fontsize=0, frameon=False, edges=True)

--> added 'pos', the PAGA positions (adata.uns['paga'])
In [9]:
sc.pl.paga(adata_ann,
    threshold=0,           
    solid_edges='connectivities_tree',
    dashed_edges='connectivities',
    layout='rt_circular',
    node_size_scale=0.5,
    node_size_power=0.9,
    max_edge_width=0.7,
    fontsize=3.5)

--> added 'pos', the PAGA positions (adata.uns['paga'])

Genes by genes

In [10]:
sc.pl.draw_graph(adata_ann, color=['Itm2a', 'Ccr9', 'Pdcd1', 'Cd27'], legend_loc='on data')
In [11]:
sc.pl.draw_graph(adata_ann, color=['Cd28', 'Slamf6', 'Sox4', 'Lef1'], legend_loc='on data')
In [12]:
sc.pl.draw_graph(adata_ann, color=['Id3', 'Tox', 'Plac8', 'Zbtb16'], legend_loc='on data')
In [13]:
sc.pl.draw_graph(adata_ann, color=['Drosha', 'Ran', 'Tcf7', 'Rorc'], legend_loc='on data')
In [14]:
sc.pl.draw_graph(adata_ann, color=['Blk', 'Izumo1r', 'Klrb1c', 'Nkg7'], legend_loc='on data')
In [15]:
sc.pl.draw_graph(adata_ann, color=['Cxcr3', 'Gzmb', 'Ifngr1', 'Il17re'], legend_loc='on data')
In [16]:
sc.pl.draw_graph(adata_ann, color=['Il1r1', 'Il18r1', 'Aqp3','Ccr2'], legend_loc='on data')
In [17]:
sc.pl.draw_graph(adata_ann, color=['Ccr6', 'Klra5', 'Klrc2', 'Klre1'], legend_loc='on data')
In [18]:
sc.pl.draw_graph(adata_ann, color=['Tbx21', 'Ccr7'], legend_loc='on data')

PAGA without Cycling and NKT1d cells

In [19]:
cell_bool = []
for x in adata_ann.obs['cell_type']:
    cell_bool = cell_bool + [x in ['NKT0','NKT17','NKT2a','NKT2b','NKT1a','NKT1b','NKT1c']]
In [20]:
list_of_cell_names = adata_ann.obs.loc[cell_bool, :].index.tolist()
adata_ann = adata_ann[list_of_cell_names, ]
adata_ann.shape
Out[20]:
(2955, 12958)
In [21]:
sc.tl.paga(adata_ann, groups='cell_type')

running PAGA
    finished: added
    'paga/connectivities', connectivities adjacency (adata.uns)
    'paga/connectivities_tree', connectivities subtree (adata.uns) (0:00:00)
In [22]:
sc.pl.paga(adata_ann, color=['cell_type'], threshold=0.05)

--> added 'pos', the PAGA positions (adata.uns['paga'])
In [23]:
sc.tl.draw_graph(adata_ann, init_pos='paga')
sc.pl.draw_graph(adata_ann, color=['cell_type'], legend_loc='on data')

drawing single-cell graph using layout 'fa'
    finished: added
    'X_draw_graph_fa', graph_drawing coordinates (adata.obsm) (0:00:16)
In [24]:
sc.pl.paga_compare(
    adata_ann, threshold=0.3, title='', right_margin=0.2, size=10, edge_width_scale=0.5,
    legend_fontsize=0, fontsize=0, frameon=False, edges=True)

--> added 'pos', the PAGA positions (adata.uns['paga'])
In [25]:
sc.pl.paga_compare(
    adata_ann, threshold=0.05, title='', right_margin=0.2, size=10, edge_width_scale=0.5,
    legend_fontsize=0, fontsize=0, frameon=False, edges=True)

--> added 'pos', the PAGA positions (adata.uns['paga'])
In [26]:
sc.pl.paga(adata_ann,
    threshold=0,           
    solid_edges='connectivities_tree',
    dashed_edges='connectivities',
    layout='rt_circular',
    node_size_scale=0.5,
    node_size_power=0.9,
    max_edge_width=0.7,
    fontsize=3.5)

--> added 'pos', the PAGA positions (adata.uns['paga'])
In [27]:
sc.pl.draw_graph(adata_ann, color=['Itm2a', 'Ccr9', 'Pdcd1', 'Cd27'], legend_loc='on data')
In [28]:
sc.pl.draw_graph(adata_ann, color=['Cd28', 'Slamf6', 'Sox4', 'Lef1'], legend_loc='on data')
In [29]:
sc.pl.draw_graph(adata_ann, color=['Id3', 'Tox', 'Plac8', 'Zbtb16'], legend_loc='on data')
In [30]:
sc.pl.draw_graph(adata_ann, color=['Drosha', 'Ran', 'Tcf7', 'Rorc'], legend_loc='on data')
In [31]:
sc.pl.draw_graph(adata_ann, color=['Blk', 'Izumo1r', 'Klrb1c', 'Nkg7'], legend_loc='on data')
In [32]:
sc.pl.draw_graph(adata_ann, color=['Cxcr3', 'Gzmb', 'Ifngr1', 'Il17re'], legend_loc='on data')
In [33]:
sc.pl.draw_graph(adata_ann, color=['Il1r1', 'Il18r1', 'Aqp3','Ccr2'], legend_loc='on data')
In [34]:
sc.pl.draw_graph(adata_ann, color=['Ccr6', 'Klra5', 'Klrc2', 'Klre1'], legend_loc='on data')
In [35]:
sc.pl.draw_graph(adata_ann, color=['Tbx21', 'Ccr7'], legend_loc='on data')

KO without Cycling and NKT1d

Back to top

In [36]:
adata_ann = sc.read_h5ad('./output/ko.ann.h5ad')
adata_raw = sc.read_h5ad('./output/ko.preprocessing.h5ad')
adata_raw.shape
Out[36]:
(4184, 13720)
In [37]:
cell_bool = []
for x in adata_ann.obs['cell_type']:
    cell_bool = cell_bool + [x in ['NKT0','NKT17','NKT2a','NKT2b','NKT1a','NKT1b','NKT1c']]
In [38]:
list_of_cell_names = adata_ann.obs.loc[cell_bool, :].index.tolist()
adata_ann = adata_ann[list_of_cell_names, ]
adata_ann.shape
Out[38]:
(2879, 13720)
In [39]:
sc.tl.paga(adata_ann, groups='cell_type')

running PAGA
    finished: added
    'paga/connectivities', connectivities adjacency (adata.uns)
    'paga/connectivities_tree', connectivities subtree (adata.uns) (0:00:00)
In [40]:
sc.pl.paga(adata_ann, color=['cell_type'], threshold=0.05)

--> added 'pos', the PAGA positions (adata.uns['paga'])
In [41]:
sc.tl.draw_graph(adata_ann, init_pos='paga')
sc.pl.draw_graph(adata_ann, color=['cell_type'], legend_loc='on data')

drawing single-cell graph using layout 'fa'
    finished: added
    'X_draw_graph_fa', graph_drawing coordinates (adata.obsm) (0:00:16)
In [42]:
sc.pl.paga_compare(
    adata_ann, threshold=0.05, title='', right_margin=0.2, size=10, edge_width_scale=0.5,
    legend_fontsize=0, fontsize=0, frameon=False, edges=True)

--> added 'pos', the PAGA positions (adata.uns['paga'])
In [43]:
sc.pl.paga_compare(
    adata_ann, threshold=0.05, title='', right_margin=0.2, size=10, edge_width_scale=0.5,
    legend_fontsize=0, fontsize=0, frameon=False, edges=True)

--> added 'pos', the PAGA positions (adata.uns['paga'])
In [44]:
sc.pl.paga(adata_ann,
    threshold=0,           
    solid_edges='connectivities_tree',
    dashed_edges='connectivities',
    layout='rt_circular',
    node_size_scale=0.5,
    node_size_power=0.9,
    max_edge_width=0.7,
    fontsize=3.5)

--> added 'pos', the PAGA positions (adata.uns['paga'])
In [45]:
sc.pl.draw_graph(adata_ann, color=['Itm2a', 'Ccr9', 'Pdcd1', 'Cd27'], legend_loc='on data')
In [46]:
sc.pl.draw_graph(adata_ann, color=['Cd28', 'Slamf6', 'Sox4', 'Lef1'], legend_loc='on data')
In [47]:
sc.pl.draw_graph(adata_ann, color=['Id3', 'Tox', 'Plac8', 'Zbtb16'], legend_loc='on data')
In [48]:
sc.pl.draw_graph(adata_ann, color=['Drosha', 'Ran', 'Tcf7', 'Rorc'], legend_loc='on data')
In [49]:
sc.pl.draw_graph(adata_ann, color=['Blk', 'Izumo1r', 'Klrb1c', 'Nkg7'], legend_loc='on data')
In [50]:
sc.pl.draw_graph(adata_ann, color=['Cxcr3', 'Gzmb', 'Ifngr1', 'Il17re'], legend_loc='on data')
In [51]:
sc.pl.draw_graph(adata_ann, color=['Il1r1', 'Il18r1', 'Aqp3','Ccr2'], legend_loc='on data')
In [52]:
sc.pl.draw_graph(adata_ann, color=['Ccr6', 'Klra5', 'Klrc2', 'Klre1'], legend_loc='on data')
In [53]:
sc.pl.draw_graph(adata_ann, color=['Tbx21', 'Ccr7'], legend_loc='on data')